Meteorological forecasts with horizontal grid spacings of 10 km or less are now routinely available to the public. Though these forecasts are increasingly precise, precision does not guarantee accuracy. Measuring the quality of these forecasts is complicated by the very nature of their variability. Root mean square errors are often large in the presence of even slight displacement errors. For this reason, high resolution forecasts may score worse than their lower resolution counterparts, even if the forecast is quite useful. This talk will focus on the qualities that make a forecast "good", and how this goodness can be measured. Many factors must be considered, such as the quality and availability of the observations, as well as the applicability of the verifying statistics. Several new verification strategies have shown promise, and examples of their application at the Naval Research Laboratory will be presented.